- Source
- All-In Podcast
- Published
- Runtime
- 55:58
- Snippets
- 21
A conversation between
Inside Trump's Science Agenda: Anti-Science Claims, Fauci's Damage, DEI & China w/ Michael Kratsios
§02
Snippets
-
It is not anti-science. I think, uh, one of the things that I'm most proud of is the release of a new report a couple weeks ago called Science, a new golden age. Um, and I think anyone who reads that report, I think what try what we will most likely see is an administration that deeply cares about the American science and technology enterprise.
Sets the administration's official self-narrative on science at the outset, against a Nature poll showing 86% of scientists backed Harris.
-
I think over the last, you know, 15 or 20 years, I think science has been deeply and deeply politicized. And we no longer are asking sort of the very hard and important questions of, you know, what is a scientific method? How should we be approaching it? Should we be questioning um, you know, some of these conclusions uh, and rather uh, you know, I think it's been dominated by kind of this this dogma. And I think it really crescendoed and peaked in in COVID where um suddenly, you know, there was a certain individual that if you didn't agree with what he said, then suddenly you were anti-science.
Articulates the administration's core critique of recent scientific institutions — that consensus became dogma — which underpins its funding overhaul.
-
Right now if you talk to sort of your your your median lobbyists for the science community, the only thing that they are fixated on is singularly is the research and development budget. If the number doesn't go up, then they haven't done their job and they haven't quote unquote supported science. And to me, that's a question. But the more important question that every scientist should be asking is, is the way that the US government spends at $200 billion in S&T funding every year, is it actually driving the best breakthroughs and biggest breakthroughs of the American people?
Reframes the science-funding debate from 'how much' to 'how well,' challenging the dominant political logic of science advocacy.
-
Senator Ted Cruz and the committee on uh at on in the Senate did did an analysis of the grants that were given during the Biden administration and they found that roughly one quarter or 25% of grants at NSF during that time period went towards DEI related quote unquote science. And if you think about that, that's an astronomical amount. That is roughly $2 billion a year times four years. That's $8 billion of science funding that went to these DI related initiatives. And that isn't science and that shouldn't be.
This is the administration's primary public justification for terminated NSF grants, and its accuracy and methodology are heavily contested.
-
The thing that is not true and what the data does not support is that it is a climate emergency. And I think one of the best examples of that is is what's happened with these um these the RCP 8.5 they call it. So these are scenarios that that climate scientists um try to estimate kind of the impact of this quote unquote climate change will have on on on the world. And this was the most extreme um scenario that had been in um in numerous national climate assessments and run by the IPCC as well. And ultimately they determined a few months ago that um they could not with any scientific integrity substantiate continuing to have this scenario play out.
Illustrates how the administration distinguishes between accepted climate science and specific modeling scenarios to justify redirecting research funding.
-
1998 14 billion 2003 27 billion 2024 47 billion so the budget has more than tripled since 1998 but there has not been a proportional rise in breakthrough treatments coming out of that funding arguably people call it heir's law it's like Moore's law in reversed it's fallen roughly 80fold since 1950 50 in terms of outcomes per dollar spent. Uh and it habs every 9 years. So every 9 years we get half as efficient as we were at getting a return on the investment we're making.
Presents the empirical case for scientific stagnation in the US — the core problem the 'New Golden Age' report is designed to address.
-
I think broadly the issue is we have not been innovative in any way on the way that we actually conduct the science the you know whether it's at NIH or at NSF or at any of our other science agencies the answer has always been let's just keep doing the same thing but add more money and hope that we get more outcomes proportion rather than asking the harder questions about are there other ways that we could be conducting the science, are there other types of scientists that could be getting the funding, could be getting the money, are there other institutions that could be getting the funding?
Identifies institutional inertia — not just money levels — as the root cause of declining research productivity.
-
One one thing that that we're going to be testing at at National Science Foundation is a concept called golden tickets where each of the evaluators are going to be given one or two or three golden tickets. And this was um initially actually tested in Denmark and some other places. And the theory is that you know the the um reviewer will then be able to unilaterally without the rest of the of the committee agreeing can make a selection of of a particular grant. So, you incentivize folks to to have more um more interesting grants.
Introduces a concrete, experimentally-tested reform to peer review designed to fund high-risk, unconventional research.
-
In venture capital, which you and I know well, you're going to have one out of 10 things work. It's a power law. That thing's worth 100x. And you want to have nine out of 10 failures because that means you're taking a lot of risk. That's really how you push the envelope and discover new things and make big breakthroughs is you got to take risk, which means failure.
Proposes importing venture capital's power-law logic directly into government science funding — a provocative and debated idea.
-
If you kind of rewind history back in in 1945, uh World War II was ending. Vanver Bush who was sort of the science adviser had my role for for FDR received a letter from the president that asked him you know what do we do with the science enterprise post World War II and Vanver Bush wrote his response that became science endless frontier which is kind of the the seinal work in the way that the US government should interface with science community... And the question now is is that model that that Bush pushed forward and and and promulgated and actually led to the the great discoveries of the last 50 years, is that still as relevant today as it was then? And our answer is it's not and we need a refresh.
Places the current reform effort in historical context, arguing that the 80-year-old Vannevar Bush framework is overdue for a fundamental rethink.
-
What has happened over the last 70 years was there's been this dramatic shift. Now roughly 70% of R&D is done by by the private sector and 30 is funded by the federal government... And the question now is is that model that that Bush pushed forward... is that still as relevant today as it was then? And our answer is it's not and we need a refresh. And that is why you wrote we wrote new golden age. And for us the first basic question we always ask ourselves is is the government spending on something that the private sector or others in the community like philanthropy are not incentivized to do.
Articulates the guiding principle for deciding what government should fund: fill the gaps private capital won't touch.
-
The alternative to an extreme is kind of what what uh what the PRC or what China is doing. And there you have sort of a a top-down directed um sort of single agency entity that that sets priorities and tries to execute. um they have been trying to essentially figure out how to do EUV lithography since since we put the export controls in 2019. No breakthrough has happened. There's nothing more important to their economy than solving that scientific problem and haven't been able to do that. So to me I feel like one of the most special features about our system is we do have people competing. This sort of free market approach to innovation is a is a huge feature of our ecosystem.
Uses China's failure on EUV lithography to argue that decentralized, competitive research systems outperform top-down directed ones.
-
I think I think the Chinese realized a few like years ago that technological and scientific leadership is the most important foundational block to economic and national security. Everything that everything that we do as a country is sort of at some in some way in my opinion rooted in scientific and technological discovery. Um we see it with what's going on in semiconductors today. The breakthroughs that that we had in transformers and other technology led to the large language models of today. literally everything that's sort of powering our economy at its core was some sort of scientific discovery.
Frames the US-China science race as fundamentally a competition for economic and national security supremacy, not just academic prestige.
-
A statistic that I I've tracked for a long time is is sort of the the percentage of USborn, you know, PhD recipients in computer science, for example. If you if you rewind the clock, you know, 30 years, I think it was, you know, 70% were American and 30% were were foreign. And now it's now it's inverted. Um, and I think to us as a country, you know, I believe that having a a a a STEM literate workforce is one of the most important superpowers we could have as a country.
The inversion of domestic vs. foreign PhD recipients in STEM is a concrete data point about America's homegrown talent pipeline crisis.
-
We used to have programs at National Science Foundation that would actually um support and encourage gifted and talented students in American high schools and middle schools. Um those were stopped for some reason. Encouraging high-erforming young students to pursue STEM in America was something that an actual decision was made of the government to to no longer fund that.
Reveals a specific and largely unreported policy reversal — defunding gifted STEM programs — that Kratsios links to DEI priorities crowding out merit-based advancement.
-
I I really do often times think about kind of what happened during COVID and and kind of what what Fouchy represented. And I think, you know, he did more disservice to science than than anyone in in modern history. you know, when he is advocating for positions like, you know, students need to be masked in schools, when um societies, you know, you know, the pediatric societies are putting out letters saying, you know, um it's okay for people to go out um and and uh collectively protest, but they're not allowed to, you know, go see their dying grandmother in a hospital. you know, no regular median American can listen to that and and then take science seriously.
Delivers the administration's sharpest indictment of COVID-era science communication as self-defeating and trust-destroying.
-
What was scary to me was how so many people without empiricism, the whole fundamental basis of science is you gather empirical data. You ask questions, you inquire, and then you go gather data and you use that data to inform your conclusion. And there wasn't a lot of empirical data that was being used to make conclusions. And everyone lined up behind these conclusions and these assumptions and it was told trust the science. And as a result, not only did it kind of destroy trust in science, but I think it brings to light questions. What happened to the scientists? Why did they all kind of line up like this? And why did everyone feel it was okay to put down people that asked questions when the basis of science is to ask questions?
Pinpoints the specific mechanism by which COVID eroded scientific credibility — enforced consensus that suppressed empirical questioning.
-
One of the statistics which was most alarming for me which uh which Jay the Jay Baracharia the director of NIH shared with me was that the median age of an intramural researcher at NIH is 71 years old.
A single striking data point — median NIH intramural researcher age of 71 — illustrates a deep generational pipeline failure in American science.
-
The idea in in science may be the same, which is you find great people. Yes. You give them significant funding. You let them decide how to spend the money rather than have some overlord that scrutinizes every dollar they're spending and every action they're taking. Gets out of their way and says, 'I trust that you're going to get somewhere. Here's a whole bunch of money, a lot more than you're asking for. Here's a whole lot of time. See you in 10 years. You'll figure something amazing out.'
Proposes a radical shift from project-based to person-based funding, a model championed by organizations like the Howard Hughes Medical Institute.
-
I think too and I again to keep harping on it. I mean honestly the the core tenant of new golden age is about thinking about how we can elevate the scientist again how the way that we fund our programs, the types of programs we do fund, the way that we structure our grants and fellowships are all centered around the scientist himself. It's not about what institution you work at or where you came from or where you grew up. It's about you yourself. Can you pursue sort of gold standard scientific work and we'll we'll support.
Crystallizes the 'New Golden Age' philosophy as scientist-first rather than institution-first — a genuine structural departure if implemented.
-
That is where we want to essentially double the scientific output of the United States by applying AI to our hardest scientific challenges and endeavors. We fundamentally believe that AI is going to be the biggest unlock to scientific discovery in the history of the world. Whether you're in material science, whether you're in chemistry, whether you're in math, whether you're in physics, you know, applying AI to your discipline is going to fundamentally change the way that you can do your science and accelerate the way that it's done.
The 'Genesis mission' — using AI to double US scientific output — is the administration's most ambitious and verifiable science claim.
§03
Synthesis
The Case for Refounding American Science
Michael Kratsios, director of the White House Office of Science and Technology Policy, argues that American science isn't broken—it's bureaucratized. The $200 billion the government spends annually on research and development produces fewer breakthroughs per dollar than it did decades ago. The solution isn't more money. It's rethinking how that money flows, who gets it, and what the government should fund in the first place.
The Efficiency Crisis in Science
The numbers are striking. The National Institutes of Health budget tripled from $14 billion in 1998 to $47 billion in 2024. Yet the return on investment has plummeted. According to what researchers call "Eroom's Law" (Moore's Law backwards), the cost to produce a new drug has doubled roughly every nine years. We fly slower than we did in the 1990s. Medical breakthroughs have stalled relative to funding increases.
This isn't because scientists are lazy or problems are inherently harder—it's because the system that distributes money hasn't evolved since Vannevar Bush designed it in 1945. Back then, the federal government funded 70% of all R&D. Private industry funded 30%. Today those numbers have flipped. Yet government funding mechanisms haven't adapted. They still treat universities as the default home for all federally funded research, impose rigid timelines (most NSF grants are locked at 18 months regardless of what the science requires), and prioritize budget size over outcome.
The Politicization Problem
Kratsios identifies a second dysfunction: the collapse of the scientific method as a cultural practice. During COVID, disagreement with specific health policy conclusions was branded as "anti-science." This wasn't a disagreement about data—it was treated as heresy. The result: the public lost faith in science itself, not because science failed, but because scientists abandoned the core principle that defines science: the willingness to question, test, and revise.
This erosion accelerated because science funding became entangled with political priorities. A Senate analysis found that roughly 25% of National Science Foundation grants during the Biden administration were explicitly tied to DEI initiatives—about $8 billion over four years. Whether one supports those initiatives, the mechanism violates basic scientific principle: grants should reward merit and ideas, not demographic boxes checked in an application.
"The most anti-science conclusion you could ever make is to say that if you don't agree with one individual, then you're anti-science."
The politicization runs deeper than grant language. It poisoned the relationship between scientists and the public. Science was treated as doctrine handed down from authority rather than as a process of inquiry open to skepticism.
The Portfolio Approach: Reforming How Money Flows
Rather than cutting budgets, Kratsios proposes restructuring them. The administration's "Science, a New Golden Age" report sketches a portfolio approach across four models:
Funding individuals instead of institutions. The Graduate Research Fellowship Program at NSF funds 2,600 of the nation's best PhD students with fully portable grants—they can take the money anywhere. This rewards talent over institutional prestige. The venture capital world figured this out decades ago: bet on great people, not ideas or places.
Experimenting with grant structures. The report proposes "golden tickets"—a mechanism tested in Denmark where peer reviewers can unilaterally approve one or two grants without committee consensus. This incentivizes bold, unconventional proposals. Under the current system, scientists propose safe projects that review committees will approve. Risk-taking looks like failure, so it gets filtered out. But venture returns follow power laws: nine out of ten bets fail, but the one that works returns 100x. Science funding structures reward caution instead.
Similarly, grants should vary in duration. Some discoveries require six months; others need five years. Locking everything to 18 months is administrative convenience, not science.
Directing universities to compete for talent. When fellowships are portable, universities have to attract researchers rather than assuming they'll stay. This introduces market discipline into academia.
Funding non-traditional institutions. Howard Hughes Medical Institute, Max Planck, Mayo Clinic—these aren't universities or government agencies. They're nimble, focused, and produce outsized breakthroughs. The government should fund the best science wherever it happens, not assume universities are the automatic home.
The China Problem
While the United States has increased R&D spending roughly threefold since 2000, China has increased spending nineteenfold. More alarming: China now publishes roughly 50% more scientific papers than the United States across nearly every field. The efficiency gap is real. Some estimates suggest China generates 2 to 10 times more output per dollar spent—partly because labor and supply chain costs are lower, but also because centralized planning allows focused deployment on specific strategic goals.
Yet Kratsios resists the temptation to solve this by centralizing American science. China's attempt to develop EUV lithography independently since 2019, despite throwing massive resources at the problem, has failed. Centralized systems excel at resource mobilization but struggle with innovation precisely because they can't generate the diversity of approaches that market competition produces.
The real Chinese advantage is focus. The government identifies priorities—AI, quantum computing, semiconductors—and directs capital accordingly. America's decentralized system generates more ideas but sometimes struggles to identify which bets matter most. The solution isn't to become China. It's to maintain competitive diversity while being intentional about which unsolved problems only government can tackle.
The Pipeline Crisis
One number haunts Kratsios: the median age of an NIH intramural researcher is 71 years old. The system has aged. And the pipeline has broken. Thirty years ago, 70% of computer science PhD recipients were American. Today that's inverted—70% are foreign-born, with most returning home or going elsewhere. The country is training the world's scientists but not retaining American talent.
This wasn't inevitable. The National Science Foundation once funded programs identifying gifted teenagers and encouraging them toward STEM. Those programs stopped—allegedly as part of broader DEI-focused policy reversals. Universities stopped using SAT scores for admissions, purportedly to reduce bias. The result: incoming students at elite institutions now require remedial math. Professors who initially supported dropping standardized tests now say they were wrong.
Kratsios argues these weren't neutral policy choices. They actively discouraged high-performing American students from pursuing science. Meanwhile, China, India, and Europe are actively recruiting talented scientists with funding, prestige, and opportunity. America is competing with one hand tied behind its back.
The Big Bets
The administration has announced several multi-year, government-scale missions: American boots on the moon by 2028; a nuclear reactor in space by 2028; a scientifically relevant quantum computer by 2028; fusion power by 2035. These are moonshot-style commitments that only government can coordinate.
But there's a tension. Quantum computing, AI, and fusion all have substantial private-sector investment. Why should government spend money here? Kratsios's answer: the private sector optimizes for commercial applications. Government should fund breakthrough science—the kind that doesn't have a clear path to profit but generates discoveries that transform entire fields. A quantum computer designed for scientific discovery will have different properties than one designed for commercial use. The government should build the former, not compete in the latter.
The administration also announced the "Genesis Mission": using AI to accelerate discovery across all scientific disciplines. This isn't about funding AI companies. It's about equipping government-funded researchers in physics, materials science, chemistry, and mathematics with AI tools to do their work better. The private sector is flooding capital into AI. The government's role is ensuring government-funded scientists aren't left behind.
The COVID Reckoning
Kratsios doesn't mince words about Fauci. He views him as having done "more disservice to science than anyone in modern history" by conflating scientific authority with political mandate—declaring mask rules and protest restrictions in the name of science, then punishing anyone who questioned those specific policies as "anti-science."
This collapsed trust. When people hear scientists say "the science is settled" and "don't ask questions," they're hearing science as dogma, not as method. And that's the opposite of science.
"If you ask one question, you're anti-science. That's the most anti-science conclusion you could ever make."
The refounding must include cultural repair: reminding people that science is a process, not a priesthood. Scientists should be comfortable with skepticism, disagreement, and revision. The moment science stops tolerating questions, it stops being science.
What Comes Next
The administration has already begun implementation through agency budget priorities and new grant mechanisms. Focused research organization funding through NSF's "XLABS" program; four-year PhDs in partnership with industry; meta-science units at NSF and NIH designed to run experiments on funding mechanisms themselves.
Whether this holds depends on politics. If a future administration reverses course, the institutional memory may vanish. But Kratsios argues the ideas are bipartisan because they're functionally sensible: if you care about scientific breakthroughs, you should fund the best scientists wherever they are, ask hard questions about whether government is deploying capital wisely, and design systems that encourage risk-taking rather than caution.
The real competition isn't about budget size. It's about which system produces more people willing to work on hard problems, more institutions capable of executing them, and more willingness to try bold ideas that might fail. On that measure, America still has an edge—if it remembers to use it.
§04
Fan-out
Questions raised
- 01 How can an administration simultaneously claim to champion science while cutting research grants?
- 02 Where is the line between healthy scientific consensus and dogma, and who gets to draw it?
- 03 What metrics should be used to evaluate whether federal R&D spending is producing genuine breakthroughs?
- 04 How did Cruz's committee define 'DEI-related science,' and was that definition methodologically sound?
- 05 Can diversity in scientific teams or research populations constitute legitimate scientific methodology?
- 06 Does retiring an extreme scenario like RCP 8.5 validate skepticism about climate urgency, or is it normal scientific refinement?
- 07 Is declining R&D productivity an inherent feature of scientific maturity, or a fixable institutional failure?
- 08 What would a genuinely experimental approach to funding science look like at the agency level?
- 09 Does empowering individual reviewers reduce groupthink or introduce greater personal bias into grant selection?
- 10 Is the venture capital model an appropriate framework for basic science, where payoff timelines are measured in decades not years?
- 11 Given that the private sector now funds 70% of R&D (vs. 30% in 1945), what should the government's unique role in science be today?
- 12 As philanthropy and focused research organizations grow, does the government need to shrink its role or redefine it?
- 13 Is China's difficulty with EUV a fair test of centralized science, given Western export controls actively blocking their progress?
- 14 If scientific leadership is foundational to national security, should US science funding be treated as a defense expenditure?
- 15 Is the reliance on foreign STEM PhD students a vulnerability or a sign of America's continuing global attractiveness?
- 16 What policy changes in K-12 education most directly affect the number of Americans pursuing STEM PhDs?
- 17 When and why exactly did NSF stop funding gifted and talented STEM programs, and what replaced them?
- 18 Is the erosion of trust in science a communication failure, an institutional failure, or both — and how do you rebuild it?
- 19 Were the inconsistencies in COVID guidance (masking, protests, hospital visits) principled exceptions or damaging contradictions?
- 20 What social and institutional pressures caused scientists to self-censor during COVID, and do those pressures still exist?
- 21 What structural features of NIH's tenure and funding systems have concentrated resources in older researchers at the expense of younger scientists?
- 22 How do you identify 'great' scientists early enough to bet on them as people rather than on their specific project proposals?
- 23 How do you evaluate individual scientific potential independent of institutional affiliation in a system built around universities?
- 24 How do you measure 'scientific output' well enough to know if it has doubled, and what would count as success for the Genesis mission?
Concepts to learn
- 01 OSTP – Office of Science and Technology Policy
- 02 Scientific method
- 03 Extramural vs. intramural research funding
- 04 Meta-science
- 05 NSF – National Science Foundation
- 06 RCP 8.5 (Representative Concentration Pathway 8.5)
- 07 Eroom's Law
- 08 Moore's Law
- 09 Meta-science units
- 10 Golden ticket grant selection
- 11 Power law distribution
- 12 Basic vs. applied research divide
- 13 Market failure in basic research
- 14 EUV lithography (Extreme Ultraviolet Lithography)
- 15 Transformer architecture
- 16 STEM pipeline
- 17 Gifted and talented education (GATE)
- 18 Empiricism
- 19 Epistemic cowardice
- 20 Intramural vs. extramural research
- 21 Investigator-based vs. project-based funding
- 22 Portable fellowships (e.g., NSF GRFP)
- 23 AI for science (AI-accelerated discovery)
References invoked
- 01 Science: A New Golden Age (OSTP report, 2025)
- 02 Anthony Fauci (implied 'certain individual')
- 03 IPCC – Intergovernmental Panel on Climate Change
- 04 Danish research council experiments with alternative peer review mechanisms
- 05 Vannevar Bush, 'Science: The Endless Frontier' (1945)
- 06 Howard Hughes Medical Institute; Max Planck Institute
- 07 US semiconductor export controls (2019, 2022)
- 08 Anthony Fauci, former NIAID Director
- 09 Jay Bhattacharya, Director of NIH
- 10 Howard Hughes Medical Institute (HHMI) investigator program
- 11 AlphaFold (DeepMind) — a precedent for AI dramatically accelerating scientific discovery (protein folding)
Mine your own.
Lode is a workbench, not a feed. Paste a YouTube URL. The model proposes a transcript, a set of quote-grounded snippets, a synthesis essay, and the fan-out. You decide what stays.